CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   tracer flow (

uma October 18, 2003 12:25

tracer flow
Hello All, I need help with the flow of a tracer in a fluid. I am using the multispecies model. I am trying to give a step input of the tracer( with a udf of step input with a velocity of 3.145 m/sec) and the other fluid flows with a velocity of 0.153m/sec at constant flow rate.Both the fluids have similar viscosities and densities. When you give a step input to a system, you have to find a bell curve concentration profile at the outlet( trying to plot the concentration of the tracer with flow time). But I am getting the profile of the tracer just as the input, I mean a perect square. Can anyone help me to know if there is some mistake in my idea. Or, is it that you can not get a bell curve, when you give a step input. I thought that when I use the model above, it may be that there is nothing much going on in the system, and so I am getting a square profile of the concentration at the outlet. So, i tried to give to give few other inlets, through which the fluid enters and then find the concentration profile of the tracer. But even then, the profile is a square. can anyone help me! Thanks a lot for taking your time. Best regards, Uma

thomas October 18, 2003 16:48

Re: tracer flow
Hi, I do not think you made a mistake in your setting. However according to your flow parameter I am not really surprise of your result. You have a Fluid 1 flowing at 0.153 m/s. The aim of a tracer is to roughly charaterize the behaviour of this fluid 1 in a finite domain ( residence time, local mixing capability ect ect ). If I follow what you say you inject a tracer (Fluid 2) at 3,145 m/s which has the same density. With such a velocity, don't you think it is this Fluid 2 which is gonna 'dominate' the flow in the domain. Therefore I am not surprise that you found such a result at the outlet if you intiate a step of Fluid 2 using this velcotiy. Have done some previous experiments using those velocities ? what are the results ? If those velocities are those you have to study, what I could suggest is to maybe just introduce your tracer not using a step, but just a Dirac impulse. What do you think about all that ?

cheers thomas ( anti-psg ;) )

uma October 19, 2003 19:56

Re: tracer flow
Hello Thomas! Thanks a lot for taking your time. I did not do any experiments for this yet. I am trying to see if the model works and then do the experiment. As you said, the velocity of the tracer must be dominating the system, so I will try to give smaller value and do the simulation. Also, by Dirac impulse, do you mean a high input value for a very less time. I usually give the tracer input as the unsteady state input and plot it as the surface integral of the concentration vs flow time at the outlet. In that case, what amount of flow time should I use for the dirac impulse at the inlet(0.2 or 1 or what is the value).I hope I am clear. Once again thanks a lot for taking you time. Best Regards, Uma

thomas October 20, 2003 04:35

Re: tracer flow
Hello, The signal you set at the inlet of your domain depends on what information your looking for ! For example you can be searching if you reactor closer than a slug reactor or than a CSTR. Personnaly, i am sorry I cannot bring you real futher information wihtout having the problem under my eyes. Just let me know what are your results using smaller tracer velocity, or shorter time injection at the inlet.

Cheers thomas

uma October 21, 2003 21:47

Re: tracer flow
Hello Thomas, I am able to get a curve instead of the step concentration profile for lower velocities. Thanks a lot for all the guidance and help. Best regards, Uma

thomas October 22, 2003 10:02

Re: tracer flow
Just for personal interest, can you describe me a little bit the type of reactor, process , or whatever thing you are trying to simulate ???? thanks. thomas

uma October 22, 2003 14:13

Re: tracer flow
Hello Thomas, I am actually doing study of the contrast in the veins. This, I guess is similar to the injection of a tracer in a vessel. I hope I have answered your question. Best Regards, Uma

thomas October 23, 2003 14:16

Re: tracer flow
HI, Your problem made me wonder about what kinda of information you could get by using a step impulse for the tracer. I am just more familiar with normal injection or dirac impulse. thomas

uma October 28, 2003 13:13

Re: tracer flow
Hello Thomas, I apologise for getting back late. I am very sorry about this. Regarding the question, in the Clinical cases, usually the contrast material is given as a step input,not an exact step(more like a trapezoidal but with very less inclination and tends to look more like a square) and not as a dirac impulse. I hope I have answered your questions. Best Regards, Uma

zwdi October 28, 2003 18:11

Re: tracer flow
Hello Uma, I am trying to simulate the traceflow also. In my case, no energy is involved. I want to use multispeices model as you did also. But I could not find this model in Fluent. I only can find SPEICES model in Fluent. Could you please give me some help? Did you use UDF for scalar transportation? Thank you very much. Best regards, Zwdi

uma October 29, 2003 11:54

Re: tracer flow
Hello Zwdi, I think if your problem is similar to mine, then you can work with the species model. This model is going to solve for all the species that u this is infact the multispecies model i think. You have to define the materials that you want to use in the model. Even I dont need the energy equation, I do not know how to eliminate it. But one way where I can get the residuals very very low, I mean a straight line with a very low value in the residuals plot. For this, after u select the species model in fluent, go to materials panel. In the mixture-type box, scroll and select mixture. Then go to Cp in the column below and change its value as constant and make the value zero. I dont know if this will help you. But you can give a try. I hope I am clear, else please let me know. And i did not use any UDF for scalar transportation. Good luck! Best Regards, Uma

All times are GMT -4. The time now is 00:33.